stratified random sampling). The important point here is that if inventory were estimated using this kind of double sampling scheme, a rather large number of soil samples analyzed for 241an (about oat 340 in Am Area 13 and 220 in GMX) would be available for estimating contours using the computer contouring program. Since there is a safetystrong linear relationship between 241am and 239-2405, at the one shot sites, this relationship could be used to transform the Am contours into 239-2405, contours within some limit of error (which could be estimated). This approach in combination with the above- mentioned techniques (using log-transformed data and more strata) has been discussed here as they might be applied to the safety-shot sites. Their applicability to nuclear sites needs to be carefully considered before sampling plans are finalized for these areas. Doubling sampling using Ge(Li) scans for 241 an has been suggested above as a method whereby more precise estimates of inventory may be obtained at safety-shot sites with no increase in costs over presently used methods. Gilbert and Eberhardt (1975) also found for Area 13 and GMX that costs could be reduced on the order of 20 to 30% using double sampling, if the precision of present estimates of inventory (Tables 4, 5, and 6 in Gilbert et az., 1975) are adequate. While Gilbert and Eberhardt did not look in detail at the remaining safetyshot sites (Area 1l and the four TTR sites), it appears that double sampling using Ge(Li) scans could be applied there also. The same general conclusions were found by Gilbert and Eberhardt to apply to a double sampling scheme wherein field FIDLER cpm readings take the place of Ge(Li) scans for 241Am in the above description of the method. Again, the method is not feasible for low-concentration strata (Stratum 1 and 2 in Area 13 and GMX). areas, For these low-level it appears there is no substitute for wet chemistry Pu analyses. An analysis approach not tried as yet by the authors is to examine the regression relationship between distance from ground zero (GZ) and 239-2405, concentrations in soil and vegetation. Since concen- trations decrease quite rapidly with distance from GZ, this suggests one might look at the feasibility of estimating Pu concentra- 239-240 tions in soil and vegetation using this variable in conjunction with Ge(Li) scans for 241An. The rate at which plutonium concentra- tions drop off also depends on compass direction, 105 suggesting this